ICMJE Authorship Criteria: 4 Requirements Explained (2026)

Introduction

Someone has to make the decision about who gets to be an author on a research paper. Often, it is up to the editors of journals to decide whose names go on the paper. Many medical journals have settled on the ICMJE authorship criteria. Under these criteria, authorship is determined by four things. First, a person had to be involved in the research. Second, that person had to be involved in the drafting of the paper or its review. Third, that person had to approve the final version of the paper. Last, that person had to agree to be accountable for all aspects of the research. If a person did only three of these, then that person is not an author. This article uses examples of common day-to-day research to explain the ICMJE authorship criteria.

What Are the ICMJE Authorship Criteria?

The International Committee of Medical Journal Editors (ICMJE) is a group of editors from major medical journals (i.e. JAMA and The New England Journal of Medicine). This group prepares and publishes guidelines for medical research, called the ICMJE Recommendations. Many journals ask their authors to follow the ICMJE Recommendations, and the latest version of the Recommendations was released in January 2026.

A published research paper gives credit. It can strengthen a job application or a residency application. It brings responsibility because the authors have to explain errant or fraudulent data. For these reasons, the ICMJE takes authorship seriously.

Every author of a research paper had to meet all four criteria, and every person who met all four had to be included as an author. People who meet all four criteria, but whose names do not appear on the research paper, are also regarded as having been authorship victimized.

The Four ICMJE Authorship Requirements Explained

An author has to meet every one of the following requirements, not just some of them.

Criterion 1: Substantial Contribution

To satisfy this criterion, individuals must have made a substantial contribution to at least one of the stages of the research: framing the research question, designing the research, collecting the data, analyzing the data, or interpreting the results.

The ICMJE does not define “substantial” with a fixed numerical threshold. In practice, the contribution should meaningfully shape the research. For example, a resident may help with sample size calculation for clinical research, while a doctor may spend several months evaluating study participants and documenting findings. A researcher who explains why one patient group responded differently contributes to interpreting the results. Routine data entry alone does not automatically establish a substantial contribution; the nature and extent of the work matter.

In contrast, a person who simply enters data into a spreadsheet, following someone else’s instructions, has helped the project but has not met this criterion.

It is better to discuss roles at the outset of a project to ensure fair judgement.

Criterion 2: Drafting or Critical Review

This criterion is related to the written report of the work, which is called the manuscript by the journals. A person meets this criterion by writing the manuscript or reviewing it critically for important intellectual content. This means reading it for scientific and logical deficiencies rather than simply grammar and/or spelling.

Earlier versions of the ICMJE guidance used the phrase “revising it critically.” Current wording “reviewing it critically” indicates that a co-author is not required to rewrite large portions of the paper. A detailed review of the paper for scientific content and clarity will suffice; however, this review may result in only minor edits to the text.

An example of a comment made by a co-author would be to point out that the discussion section does not address an additional variable that may account for the findings, or that the conclusion makes claims beyond what is supported by the data. These comments enhance the science being reported, thus both are considered critical reviews. On the other hand, copyediting (e.g., correcting spelling and grammar errors, formatting references) is a different type of activity. Although copyediting enhances the readability of a paper, it alone does not satisfy the criteria for a critical review.

Criterion 3: Final Approval of the Manuscript

The last condition to be met by each of the authors is their approval of the final version of the manuscript that will ultimately be published. An agreement on either the study design or an early version of the manuscript does not meet this requirement.

The paper changes substantially prior to being published. Many times these changes are due to the peer review process. Peer review is when outside researchers (experts) are chosen to review the manuscript and make recommendations regarding changes/analyses. As such, when the paper is accepted, the methodology, findings and/or conclusions could differ greatly from the original version.

For that reason, the author that is submitting the paper needs to submit the final copy to all of his/her co-authors for confirmation of their approval. Also, if there are major revisions made after peer review, then all co-authors must approve those revisions. An author approving an earlier draft does not imply that he/she agrees with the content of the final version.

Criterion 4: Accountability for the Work

The final section deals with the issue of responsibility. All of the authors have agreed to take ownership of the entire research project. Therefore, if there are questions concerning any aspect of the accuracy or truthfulness of the research, then all of the authors must ensure that those questions are investigated and resolved.

Every author does not have to be an expert in all aspects of the research project. For example, a doctor on a research team is not expected to be able to explain how to perform every single statistical analysis. However, every author should be responsible for their own portion of the study; know who performed the remaining portions of the study; and assume that all of their co-authors performed their work honestly.

Imagine two years after the article has been published, a journal inquires about a questionable value found in a table within the article. All authors are expected to identify which individual was responsible for managing the data associated with the question and then work together to resolve the issue.

Who Qualifies for Authorship? Examples From Medical Research

Job titles don’t decide authorship. What a person actually did decides it. That’s why the same role can lead to authorship on one paper and only a thank-you on another.

Contributor	Can be an author if they	Cannot be an author if they only
Research supervisor	Helped plan the study or explain the results, and meets criteria 2-4	Headed the department or gave general guidance
Data collector	Did a large share of the data collection, and meets criteria 2-4	Collected data but did not review or approve the paper
Statistician (data and numbers expert)	Helped plan the analysis and explain the results, and meets criteria 2-4	Ran one test when asked, with no other input
Medical student	Made a substantial contribution, and meets criteria 2-4	Searched for articles or entered data
Language editor	Rarely qualifies through editing alone	Fixed grammar, style, or formatting

Being senior doesn’t make you an author, and being junior doesn’t stop you from becoming one. A medical student can even be the first author if they did most of the work and meet all four criteria.

Who Does Not Qualify as an Author?

Some kinds of help are valuable but aren’t enough, on their own, to make someone an author. The ICMJE lists four: finding the money to pay for the study, general supervision of a research group, administrative help such as paperwork and scheduling, and writing or editing help such as grammar checks and proofreading.

These people should still be thanked in the acknowledgments. Say what each person did, for example “helped recruit patients” or “edited the paper for language.” If many people helped in the same way, such as doctors who enrolled patients at different hospitals, you can thank them together as a group.

Ask for written permission before you name anyone. Readers may assume that anyone named in a paper agrees with its results. So each person should get to decide whether they want their name there.

Gift, Guest, and Ghost Authorship: What’s the Difference?

These three terms describe unfair authorship. Each one gives readers a false picture of who did the work.

Gift authorship, also called honorary authorship, means adding someone who didn’t earn it, usually as a favor. The classic example is putting the department head’s name on every paper from the unit, whether or not they ever read it.

Guest authorship is similar, but the reason is different. Here, a team adds a famous researcher to make the paper look more impressive. A junior team might do this hoping the journal will be more likely to accept their work.

Ghost authorship is the reverse. Someone who did important work is hidden completely. They aren’t listed as an author, and they aren’t thanked either. A well-known example is a paper funded by a drug company and written by a writer the company paid, whose role is never mentioned. This can hide a conflict of interest, meaning a money or personal reason that could affect how the results are presented.

Authorship Order, First Author, and Corresponding Author

The four criteria tell you who is an author. They don’t tell you what order the names go in. The ICMJE leaves that choice to the authors. They should agree on it together and be ready to explain how they decided.

In medical papers, the first author is usually the person who did the most work. The last spot often goes to the senior researcher who led the project. These are common habits, not fixed rules, and other fields do it differently.

The corresponding author is the person who deals with the journal. They send in the paper, answer the journal’s questions, and reply to readers after publication. It’s a practical job, not a sign of rank, and this person can appear anywhere in the list of names.

ICMJE vs CRediT: What’s the Difference?

CRediT stands for Contributor Roles Taxonomy. A taxonomy is just a system for sorting things into groups. CRediT gives 14 standard labels for the jobs people do in a research project. A few examples are conceptualization (coming up with the main idea), data curation (organizing and cleaning the data), and formal analysis (doing the statistics). Many journals ask authors to list their CRediT roles when they submit a paper.

The two systems do different jobs. ICMJE decides who gets to be an author. CRediT describes what each person did.

So having a CRediT role doesn’t automatically make someone an author. A person who organized the data might still not meet all four ICMJE criteria. In that case, their work is described in the acknowledgments instead.

Can ChatGPT or Other AI Tools Be Listed as Authors?

No. The ICMJE says AI tools like ChatGPT should not be listed or cited as authors. It’s easy. Since a tool cannot be responsible for the work done with it, it cannot fulfill the fourth requirement.

You should also let us know when you use artificial intelligence (AI) in addition to using it. An example would be if you wrote your manuscript with the help of AI, then acknowledge the assistance in your acknowledgment section. If you used AI to gather, analyze, and/or create figures from your data, then state that in your methods section so we understand how you did your research.

Human authors remain responsible for the accuracy of their manuscript, including any material produced with AI assistance. AI tools can generate references that look authentic but do not exist, so verify each reference against its original source. Our guide to artificial intelligence in scientific writing explores its uses, limitations, and ethical concerns.

How to Avoid and Handle Conflicts over Authorship.

Most authorship disputes begin when teams leave the discussion until the work is already finished. Agreeing on contributions early and keeping a written record can help prevent misunderstandings. For projects involving human participants or identifiable data, ethics review is another issue to address during planning. Our guide explains how to get IRB approval for research. Authorship decisions and ethics review serve different purposes, so address each separately.

1. When you initially develop your research proposal, agree on who will be listed as authors and the order in which they appear. You may alter this at a later time.
2. Keep a written record of which person performed each task.
3. When people come in late, leave early, or have to do more work than expected, update the roster.
4. In international projects, use local researchers who qualify to work with you on your research. It is important to use local researchers in the location where you obtained the data. The ICMJE advises that you do not exclude these researchers from your research.
5. 5.Before submitting your work, make sure that each author has agreed upon the last list of names in the exact order they will appear in the paper.

If the dispute persists, the team must attempt to resolve the issue as a preliminary step. Journal editors do not have authority over deciding which individuals deserve to be listed as authors. In the event that the team cannot reach consensus, the university or hospital responsible for carrying out the research must review the matter.

In the event that the authors of a paper are changed after submission of the manuscript, most journals will require two conditions; consent of every author (including those added/removed) and the purpose behind the change.

ICMJE Authorship Checklist Before Manuscript Submission

Before you send your paper to a journal, check that:

• Each author made a substantial contribution, and it’s written down.
• Each author wrote the paper or reviewed it carefully.
• Each author approved the final version.
• Each author agreed to take responsibility for the work.
• People who helped are thanked with their permission, and any AI use is reported.

Then read your target journal’s instructions for authors. Check whether it requires a contributor statement, signed authorship forms, or additional disclosures. If you are new to manuscript preparation, our guide to how online research courses support writing and publication explains how structured training can help you develop these skills.

Frequently Asked Questions About ICMJE Authorship

Q1. Do all authors have to meet all four ICMJE criteria?

Yes. Every author must meet all four. Anyone who meets fewer should be thanked in the acknowledgments.

Q2. Can a supervisor be an author?

Yes, if the supervisor meets all four criteria. General guidance alone isn’t enough.

Q3. Does data collection qualify someone for authorship?

A large share of the data collection can meet the first criterion. The person still has to meet the other three.

Q4. Can an author be removed after manuscript submission?

Usually, yes. Most journals need written agreement from all authors, including the person being removed, and a reason for the change.

Q5. Who resolves authorship disputes?

The research team should try first. If they can’t agree, the university or hospital where the work was done should look into it. Journal editors don’t settle these disputes.

Final Thoughts: Apply the ICMJE Criteria Before Submission

Authorship problems are much easier to prevent than to fix. Use the four-criteria rule from day one, write down who did what, and check your journal’s rules before you submit.

Need help preparing your manuscript? The American Academy of Research and Academics supports researchers with research writing, publication ethics, and journal submission. Get in touch to make your next paper ready for submission.

Disclaimer:

Articles published by American Academy of Research & Academics are prepared by our team using information from direct experience, publicly available resources, and educational references. AI tools may be used to assist with drafting, proofreading, and formatting; however, all content undergoes review and approval before publication.
The information provided is intended for educational purposes only. Requirements, policies, and processes may change over time. Readers should consult official sources for the most current information.